IMPROVE - Interagency Monitoring of Protected Visual Environments

RHR2 -  Regional Haze Rule Version 2, Using the new IMPROVE algorithm to calculate light extinction

The IMPROVE RHR2 Program contains a version of the IMPROVE Aerosol data with major aerosol species, additional substituted daily data values, as well as data aggregations.  Detailed descriptions of the data substitution and aggregation algorithms can be found in the RHR Guidance Documents.  General descriptions are provided below.  Important difference between the IMPROVE RHR2 Aerosol data and the IMPROVE Aerosol data in the IMPROVE Raw Program on VIEWS are that the RHR data contain substituted values and that all calculated values use the new IMPROVE algorithm to calculate light extinction and its components from the raw aerosol data.

• RHR Guidance Documents
• IMPROVE RHR data products (Annual Summary)
• New Algorithm

Raw IMPROVE data flags:

• RHR program data are calculated from raw IMPROVE data with the following native flags: "NM", "LF", "RF", "AA", "AP", "SW", "SP", "NR","CG".  These flags can be found in the Flag1 column of the raw IMPROVE aerosol data.
• To determine whether a daily value contains a patched or substituted value pull the native flag #4.  Flag4 values greater than one indicate the value is or is comprised of one or more patched (or substituted) value.
• To determine to which aerosol extinction group belongs pull the native flag #5.  Flag5 values of 90 indicate the daily value belongs to the Group 90, or worst 20% aerosol extinction days.  Flag5 values of 10 indicate the daily value belongs to the Group 10, or best 20% aerosol extinction days.  Intermediate group values of Group 30, Group 50, and Group 70 are also flagged.

Parameters:

• The RHR dataset includes only major parameters, generally those required to calculate visibility metrics such as aerosol extinction, deciview, and standard visual range.
• In the VIEWS RHR2 dataset, NO2_bext is set to zero.
• The algorithm for separating ammSO4, ammNO3 and OMC into small and large fractions can be found in the New Algorithm documentation.

Metadata Flags:

VIEWS Flag:	Description:	Detailed Description:
Flag1	not used
Flag2	Number of years	Applies to multi-year aggregations, such as 3-year and 5-year rolling averages.  The number of annual averages used to form the multi-year mean.
Flag3	Number of daily values	Applies to single-year aggregations.  The number of daily values in the annual mean.
Flag4	Number of substituted values	Applies to daily values.  The number of substituted values.
Flag5	aerosol extinction group	Applies to daily values.

Note: For the IRHR program, Flag 5 can have a value of 10, 30, 50, 70, 90, or #.  If a daily value has a Native Flag 5 value of 90 (indicating a group 90 value), then this value is one of the RHR prescribed Worst 20% visibility days.   If the flag 5 is #, then the year had insufficient data to form aerosol bext groups according the the RHR guidelines and no group values are assigned for that year.

Substituted Values:

In order to calculate chemical or 'aerosol extinction' a given daily sample from an IMPROVE aerosol monitoring location must have reported values for the parameters ammSO4f (ammonium sulfate), ammNO3f (ammonium nitrate), OMCf (organic carbon mass), ECf (elemental carbon) SOILf (soil), and CM (coarse mass).  If any of these six parameters are missing the Guidance specifies procedures to substitute the missing value.  Details of the substitution methodology and criteria are given in Section 2 of the Tracking Progress Guidance document.  In the VIEWS database any of the six aforementioned parameters measured on a daily basis and having a Flag4 value > 0 indicates a patched value.  Daily reconstructed parameters, such as RCFM or aerosol extinction, can containing one or more substituted value have Flag4 values indicating the total number of patched values included in the reconstructed parameter.

Data Aggregations:

1. Data aggregations are comprised of samples collected at one monitoring site over the course of more than one sampling day.
2. Aggregated data are not currently available from the Query Wizard.  These data can currently be obtained from text files in the Annual Summary Summary Data section or viewed in tabular and graphical format from the various Annual Summary sections.
3. All aggregated visibility metrics in VIEWS are derived by first calculating the daily metric (such as daily deciview or daily SVR) then averaging the daily metrics to obtain the aggregated value.  For example, the average deciview value on the Worst 20% sampling days is calculated by taking the mean of the daily Worst 20% deciview values.  Because the decivew and SVR algorithms are not commutative their aggregated values in VIEWS cannot be reproduced directly from the mean aerosol extinction value for the same data aggregation.  That is, the the mean deciview value for the Worst 20% sampling days is not necessarily equal to the the mean extinction value on the worst 20% days converted to deciview.

Important Notice for IMPROVE RHR data:

Over the years the IMPROVE program has undergone a number of sampling and filter analysis changes to
reduce biases and uncertainty and improve measurement sensitivities. These changes are documented in
the quality control section of the IMPROVE website:

http://vista.cira.colostate.edu/improve/Data/QA_QC/qa_qc_Branch.htm

Two significant changes which can influence any trend work involved the fine sulfur mass and nitrate
ion measurements. The sulfur measurement from the Module A Teflon filter can be underestimated by over
a factor of 2. This bias occurred at only some Eastern US sites primarily during the summers of 1992-95
on days with the highest sulfur loadings coinciding with high humidity. This problem was rectified by
increasing the size of the Teflon filters at affected sites (Table 1). It is recommended that data analysts
use the sulfate ion measurement from the Module B filter instead of the Sulfur mass measurement from
module A for time periods before the Teflon filter size was increased at the sites listed below (Table 1). 
In the IMPROVE RHR data set this substitution has been made for all parameters that include ammonium sulfate
in their calculation.  In the IMPROVE RHR dataset, this substitution was not made at sites with no evidence of
sulfur underestimation.

Table 1. Monitoring sites and dates when the Teflon filter was increased from 2.2 sq. cm to 3.5 sq. cm :

Apr 19, 1995 - DOSO1,GRSM1,SHEN1,JEFF1
Apr 22, 1995 - SAGO1
Apr 26, 1995 - ACAD1,LYBR1,SHRO1,MACA1
May 03, 1995 - SIPS1,UPBU1
May 10, 1995 - BRIG1
May 27, 1998 - SEQU1
Jun 03, 1998 - PUSO1
Jul 29, 1998 - MOOS1
Aug 04, 1998 - CHAS1,OKEF1,ROMA1,EVER1
Apr 17, 2001 - CACR1
Apr 25, 2001 - HEGL1,GUMO1,BIBE1
May 16, 2001 - WIMO1

Note: there is no evidence of S underestimation at BIBE1, GUMO1, MOOS1, SAGO1 and SEQU1.

For detailed information on the bias and recommendations see the QA/QC issues page on the IMPROVE website at:

http://vista.cira.colostate.edu/improve/Data/QA_QC/issues.htm

A temporal bias also exist in the IMPROVE nitrate ion data. For a period after  6/1996 The nitrate data have smaller
concentrations and variability.  This is most pronounced in the Eastern US and along the West Coast. The
change in the nitrate data coincides with the addition of glycerin to the denuders for the Module B sampler.
However, a causality has not yet been established and is currently being studied.   In the IMPROVE RHR dataset the
the measured particle nitrate value is used to calculate ammonium nitrate.